Journal Home page : www.jeb.co.in « E-mail : [email protected] JEB ISSN: 0254-8704 Journal of Environmental Biology CODEN: JEBIDP

Effect of pesticides used in banana and pineapple plantations on aquatic ecosystems in Costa Rica

Noël J. Diepens1,2 *, Sascha Pfennig 1, Paul J. Van den Brink2,3 , Jonas S. Gunnarsson4 , Clemens Ruepert 1 andLuisa E. Castillo 1 1Central American Institute for Studies on Toxic Substances (IRET), Universidad Nacional, Heredia, 83-3000, Costa Rica 2Department of Aquatic Ecology and Water Quality Management, Wageningen University, Wageningen University and Research Centre, P.O. Box 47, 6700 AA Wageningen, the Netherlands 3Alterra, Wageningen University and Research Centre. P.O. Box 47, 6700 AA Wageningen, the Netherlands 4Department of Ecology, Environment and Plant Sciences, Stockholm University, Stockholm, SE-106 91, Sweden *Corresponding Author E-mail: [email protected]

Abstract Current knowledge on fate and effect of agricultural pesticides comes is mainly from temperate ecosystems. More studies are needed in tropical systems in order to assess contamination risks to non- target endemic tropical species from the extensive use of pesticides e.g. in banana and pineapple plantations. In this study, acute laboratory toxicity tests with organophosphate pesticides ethoprophos and chlorpyrifos were conducted on two Costa Rican species, cladoceranDaphnia ambigua and fish Parachromis dovii. Tests showed that chlorpyrifos was more toxic than ethoprophos toD. ambigua and P. Publication Info doviiand that D. ambigua was also more sensitive than P. dovii to both pesticides. Additionally, bioassays Paper received: were performed by exposingD. magna and P. dovii to contaminated water collected from the field. 27 April 2013 Chemical analyses of field water revealed that fungicides were generally the most frequent pesticide group found, followed by insecticides/nematicides and herbicides. The bioassays and values obtained from the literature confirmed thatD. magna was more sensitive to pesticide contamination thanP. dovii and that D. Revised received: ambiguawas more sensitive than D. magna , suggesting that the native cladoceran is a more suitable test 24 June 2013 species than its temperate counterpart. Species sensitivity distributions showed no significant difference in sensitivity between tropical and temperate fish and the arthropod species exposed to chlorpyrifos in this Accepted: study. Choline esterase activity (ChE) was measured inP. dovii in laboratory tests in order to assess the 05 September 2013 applicability of this biomarker. ChE inhibition inP. dovii was observed in the laboratory at levels below the

LC10 of both ethoprophos and chlorpyrifos, confirming that ChE is an efficient biomarker of exposure. Both indigenous Costa Rican species used in this study were found to be suitable standard tropical test species. Further studies are needed to investigate how protective the safe environmental concentrations, derived

from LC50 of native tropical species, are for protecting tropical aquatic natural communities. Key words Acute toxicity, Bioassays, Chlorpyrifos, ChE inhibition, Ethoprophos, Tropical aquatic ecosystems

Introduction REPCar, 2011; Castilloet al., 2012; Bravo et al., 2013). The most important group of pesticides used are fungicides, followed by a Costa Rica is the second largest banana producer in the wide range of insecticides, nematicides and herbicides (see world. Commercially grown bananas are among the world's most (Castilloet al., 1997; Ramírez, 2009) for a review). Currently, pesticide intensive crops. Approximately, 45 kg of active pineapple and rice are increasing their production area. ingredient (a.i.) per hectare per year are used in the Costa Rican Pineapple production uses an average of 30 kg of a.i. of banana plantations (Castilloet al., 1997; Ramírez,Online 2010; pesticides per hectare per yearCopy and currently the crop is grown

© Triveni Enterprises, Lucknow (India) Journal of Environmental Biology, Vol.35 Special issue, 73-84, January 2014 74 N.J. Diepens et al. over 45000 ha (Castilloet al., 2012). This together with the 20% or more is well accepted as indicator of exposure to OPs banana plantation, these crops potentially brings along more (Varóet al. , 2007). environmental pollution and health risks (Barrazaet al. , 2011). The river Rio Madre de Dios (RMD) in the Atlantic Two widely used organophosphate (OP) pesticides are lowlands of Limon, Costa Rica (Fig. 1) receives drainage water chlorpyrifos and ethoprophos. Plastic bags that are used to from banana plantations, pineapple and rice fields before it flows protect banana bunches are usually impregnated with into a coastal lagoon, which is part of a protected coastal wetland chlorpyrifos (1% mass concentration equals 0.7 kg a.i. ha-1 y -1 with a large biodiversity. Due to the high rain precipitation in this (Matlock Jr and de la Cruz, 2002). Ethoprophos is one of the most region, with an average of 3.6 m per year (Waylenet al., 1996), commonly used nematicides to protect the root systems of the intensive use of pesticides and fertilizers in these extensive banana and pineapple plants. In banana plantations, it is applied agricultural systems may reach the RMD river and its lagoon a maximum of three times a year (16-21 kg a.i. ha-1 y -1 ; (Matlock Jr trough run-off and endanger the non-target fauna such as fish and and de la Cruz, 2002)). Together with carbamate pesticides, OPs invertebrates. Several massive fish kills have been observed in are neurotoxins, which act by inhibiting the cholinesterase these water courses, especially after heavy rains and are enzymes (ChE) acetylcholinesterase (AChE) and butyryl cholin believed to be caused by pesticide run-off. Generally, fish species esterase (BChE) in the nervous system. Inhibition of ChE has such as guapote (P.dovii ), black surfperch ( Embiotoca jacksoni ), been widely used to assess the effects of anticholinesterase tilapia (Oreochromis niloticus ) and other cichlid species have compounds on wildlife (Thompson, 1999). In fish, inhibition of been found among the dead fish (Pais, 2009; Pais, 2010).

Online Copy

Fig. 1 : Sampling sites in watershed Rio Madre de Dios (14758 ha), Limon, Costa Rica. RMD is Rio Madre de Dios and LMD is Laguna Madre de Dios. RMD control was used as control site and together with Bridge RMD, Canal Pama, LMD X RMD they were used for the bioassays and ChE analyses. CañoAzul, RMD X Freeman, LMD mid and LMD x Pacuare were used for water quality and pesticide residue analyses only Effect of pesticides on aquatic ecosystems in Costa Rica 75

Besides acute mortality, pesticides may also cause reproduction A temperate and tropical species sensitivity distribution failure and other chronic effects, which could lead to a decline in calculation was used to evaluate to what extent these tropical fish populations. Furthermore, pesticide pollution may result in species are suitable for toxicity testing and environmental risk acute and chronic effects on macroinvertebrates (Castillo et al., assessment in tropical regions. 2000), which could lead to changes in the benthic community structure and species richness (Pringle and Ramírez, 1998; Materials and Methods Castilloet al., 2006), and in turn change the food web structure. Test species : Two cladoceran species (Crustacea) tropical Despite the high use of pesticides in Costa Rica and other Daphnia ambigua(Scourfield, 1947) and temperate Daphnia tropical countries and their possible environmental impact, there magna (Straus, 1820) and the guapote fish (Cyclidae) is still relatively little knowledge about the fate and toxicity of Parachromis dovii (Günther, 1864) were used as test organisms. pesticides in tropical aquatic ecosystems compared to temperate The fish were cultivated and kept under a natural light : dark cycle systems (Castilloet al., 1997; Lacher and Goldstein, 1997; (approximately 12:12 hr) in the laboratory at the Biology Mortensenet al., 1998; Pringle and Ramírez, 1998; Robinson et Department of the National University (UNA), Heredia, Costa al., 1999; Matlock Jr and de la Cruz, 2002; Kwoket al., 2007; Rica. Aged tap water (24 hr) with a temperature of 18-25°C was Daam and Van den Brink, 2010). For instance, Maltbyet al. (2005) used in the fish tanks. The guapote fish were fed a mixture of fish, collected toxicity data from different databases and only 11% of soya, cornflour, shrimps and soya oil, supplemented with vitamins the species originated from tropical regions, while more than 50% and minerals. Fry were also fed withArtemiasalina and cultivated were temperate species.There is an on-going debate on whether zooplankton. After 12 to 18 days, the fish were transferred for species sensitive distributions (SSDs) derived from temperate acclimatization for at least 24hr to the laboratory of toxicity benchmark values (i.e. LC50 or NOEC values) can be ecotoxicological studies (ECOTOX) of IRET (Central American used to assess the contamination risks for tropical species as well Institute for Studies on Toxic Substances), UNA Universidad or reversely if it is necessary to calculate benchmark values only Nacional, Heredia, Costa Rica. They had a mean length of 7.3 ± based on native tropical species. 0.5 mm and a weight of 8.8 ± 1.0 mg.D. ambigua and D. magna were cultured in laboratory of ECOTOX. The crustaceans were A few recent studies have found no significant difference kept in glass beakers of 200 ml in a temperature controlled room between the slope of SSD curves based on toxicity data for (18-22°C) with a light : dark cycle of 16:8 hr. They were fed yeast, -1 2+ -1 -1 tropical and temperate species, suggesting that SSDs obtained cereal leaves, tetramin (0.5 ml l ), Se (5.2 ml l ), B12 (0.095 ml l ), based on temperate species could be applied to protect tropical Chlorellasp. (0.275 ml l-1 ) and Selenastrum sp. (0.57 ml l-1 ). Only species (Dyeret al., 1997; Maltby et al., 2005; Kwok et al., 2007; neonates from adults that already had one or more clutches were Daam and Van den Brink, 2010; Ricoet al., 2010). In line with this, used for the tests. Neonates between 24 and 72 hr old were taken Daamet al. (2008) found no significant difference in species for acclimatization (<24 hr) and used for testing. sensitivity at the community level for chlorpyrifos between temperate and tropical aquatic ecosystems. For some pesticides, Chemicals : The tested pesticides were chlorpyrifos (CAS nr. including chlorpyrifos, data are available for the comparison of 2921-88-2, Dr. Ehrenstorfer, purity 98.4 ± 0.5%) and ethoprophos temperate and tropical species. For other pesticides such as (CAS nr. 13194-48-4, Dr. Ehrenstorfer, purity 91%).All tests were ethoprophos, however, this is not the case, and therefore conducted in the ECOTOX laboratory. Stock solutions of the temperate data must be used to perform an ecological risk pesticides were made by diluting chlorpyrifos in acetone (1:1) and assessment for tropical species. ethoprophos in deionized water from a Milli-Q water purification system (Millipore). Test concentrations for fish were prepared by diluting the stock solution with carbon filtered, UV treated water. The aim of present study was to understand the effects of For daphnids, hard reconstituted water was used, which was pesticides in tropical aquatic ecosystems by conducting -1 -1 prepared by adding KCl (0.008 g l ), MgSO4 (0.12 g l ), standardized laboratory acute toxicity tests with chlorpyrifos and -1 -1 CaSO4 .H 2O (0.12 g l ) and NaHCO3 (0.192 g l ) to deionized water ethoprophos on two Costa Rican indigenouscladoceran species, -1 (Dutka, 1989), with a conductivity of 3.96 ms cm and a pH of 8. Daphnia ambiguaand Parachromis dovii . Furthermore, water For control, same water was used as for the dilutions. The samples were collected in the field to measure pesticide residue maximum amount of acetone used in the concentration ranges levels in rivers and creeks receiving agricultural runoff. Standard was also added to the acetone control. toxicity bioassays were performed underOnline laboratory conditions Copy using these field collected water samples onDaphnia magna and A range finding test was performed with fish to determine P. dovii. ChE inhibition assays were performed in the laboratory the concentration range of chlorpyrifos for the final test. The onP. dovii to evaluate the use this biomarker for exposure concentration ranges for the other test were based on earlier assessment. Furthermore, we compared the differences in experiences at the ECOTOX laboratory or by using the geometric sensitivity of these two tropical species (fish and crustacean) to mean of LC50 values present in the EPA database their temperate counterpartsOncorhynchus mykiss and D. magna . (www.epa.gov/ecotox). 76 N.J. Diepens et al.

Acute toxicity test set up: All tests were performed in triplicate carried out with water from the four sites described above, P.dovii (n=3) with 10 individuals per treatment and performed according (72 hr, semi static with a 75% water exchange after every 24 hr) to adapted standard protocols of OECD (1992) and conducted at andD. magna (48 hr, static) were used. D. magna was used an ambient temperature (24.9 ± 1.2°C) and a light:dark cycle of instead ofD. ambigua due to availability at that moment. For both 12:12 hr. Crustaceans were exposed in 150 ml medium water in a species, ten individuals were exposed to 200 ml of field water glass beaker covered with parafilm, without aeration and food. obtained from each site. Besides the field samples and the field Ten fish were put in each beaker glass with 200 ml of aerated control, a negative control (n=3) with carbon-filtered, UV treated water and were not fed during the tests. Each beaker was water for fish and reconstituted water forDaphnia (as described continuously aerated during the tests using air pumps connected above) was also included in the experimental setup. The tests to a glass Pasteur pipette.After two days, 150 ml of test water was were conducted at ambient temperature (24.7 ± 0.9°C). Mortality replaced by new stock solution. Oxygen concentration was recorded every 24 hr forDaphnia and fish. (Wissenschaftlich-Technische Werkstätten (WTW) GmbH Oxi325), pH (Corning pHmeter 220), conductivity (WTW cond Cholinesterase assay : Surviving fish from toxicity tests and 315i), water temperature (WTW cond315i) and air temperature bioassays were analysed for cholinesterase (ChE) inhibition. At (Digital thermometer Cresta) were recorded at 24, 48, 72 and 96 the end of each toxicity tests, ten fish were randomly chosen from hr. This was done for one randomly picked replicate of each three replicates of each test concentration and from each of the concentration. Water samples of the stock solutions and of one bioassay tests. When the number of surviving fish was less than random replicate per concentration at the end of each test were ten, remaining fish were used.The fish were rinsed with deionised taken for pesticide concentration analyses to the chemical water and kept frozen individually in Eppendorf tubes at -18°C. laboratory (LAREP) of IRET,Heredia, Costa Rica. For control and Whole fish samples were thawed and homogenized in 1 ml ice- acetone control, only one replicate of each was collected for cold phosphate buffer (0.1 M, pH 7.2) using a Branson SLPt pesticide analysis. sonifier. Samples were then centrifuged at room temperature (Eppendorf 5415c, 10,000 rpm, 5 min). Supernatants were Mortality was monitored every 24 hr and was defined for transferred to new Eppendorf tubes and stored at -18°C until fish as the lack of gill movement. ADaphnia individual was further analysis. Protein concentration in the samples were considered dead if the animal did not visually move after one determined in triplicate by the Bradford method (Bradford, 1976) minute of stimulation, using a magnifying lens. Dead animals adapted to microplate (Bio-Rad Laboratories 2005). Bovineg - were removed from the test immediately. globulin was used as standard. Dilutions were made to normalize protein content to approximately 0.5 mg ml-1 . Dye reagent and Environmental sampling : Eight sites within the watershed Rio protein standard were purchased from Bio-Rad. The ChE activity Madre de Dios located in the Caribbean lowlands of Costa Rica was determined at room temperature (25°C) by the method of downstream from large banana and pineapple plantations (Fig. 1) Ellmanet al., (1961) adapted to a microplate (Lopes et al., 1996; were sampled during rainy season (8-10-2009). One sample of Menaet al., 2012). Acetylthiocholine (Sigma-Aldrich) hydrolysis surface water was collected from each site for pesticide analysis. was measured as an increase of absorbance at 414 nm caused The water samples were collected in a pre-washed 4 l capacity by the reaction of thiocholine with DTNB (Sigma-Aldrich) to amber glass bottles, which were rinsed once with surface water. produce yellow 5-thio-2-nitro-benzoic acid anion. Absorbance At four sites (Fig. 1), additional water samples were taken for was measured at an interval of 2.5 min for 15 min. For the laboratory bioassays. The bottles were kept on ice during their determination of ChE activity, change in absorbance between 5 transport to the laboratory (approx. half a day). A site situated and 10 min was calculated. A Thermo Multiskan Ascent upstream of most of the plantations was assumed to be relatively microplate reader was used for both enzymatic and protein unpolluted and therefore chosen as a reference. The site near the determinations. ChE activities were calculated as nmol of bridge over the "Rio Madre de Dios" (Bridge RMD) was chosen as hydrolyzed substrate per min and mg of protein. The results were a relative uncontaminated site but was surrounded by plantations expressed as units (U) per mg of protein (U = n mol min-1 ). and used as an experimental treatment for the bioassay. Another site, the "Canal Pama" channel, collects drainage water from Pesticide analyses : For determination of pesticides various banana plantations and was chosen for the bioassay concentration, water was extracted by solid phase extraction because of high fish mortality that has been reported from this (SPE) and pesticides were analyzed by gas chromatography with site. The last site of four selected sitesOnline for the bioassays was mass detectorCopy (GC-MS) and by liquid chromatography with located at the entrance of the river into the lagoon (RMD X LMD). photodiode array detector (LC-PDA). In the field samples, 16 At all sites oxygen concentration (WTW Oxi325), pH (WTW pesticides were analyzed: ametryn, bromacil, carbofuran, pH320), conductivity (WTW cond 315i) and water temperature carbaryl, chlorothalonil, chlorpyrifos, diazinon, difenoconazole, (WTW cond 315i) were measured. diuron, epoxiconazole, ethoprophos, fenamiphos, hexazinone, metalaxyl, terbufos, triadimenol. They were arranged in five Bioassays : For the bioassays (n=3), i.e. laboratory toxicity tests pesticide groups (insecticides, nematicides, herbicides, Effect of pesticides on aquatic ecosystems in Costa Rica 77 fungicides and insecticides/ nematicides) based on Tomlin (chlorpyrifos: 350.6 g mol-1 and ethoprophos: 242.3 g mol-1 (2003). The SPE cartridges (200 mg ENV+/6 ml Isolute or C18 (Tomlin, 2003)). All calculations were done using nominal 500mg 3ml-1 ) were previously conditioned with ethyl acetate (5 ml, concentrations. Adjustments for control mortality were made by analytical grade of Fluka Pestanal), methanol (10 ml analytical the program. The following formula, modified from Schroer et al. grade of J.T. Baker hplc solvent), deionized water (10 ml) and (2004), was used to construct a dose response curve. methanol (5 ml). The water samples (1l) were passed through the 1-c cartridges, than they were dried by centrifugation (2 min at 5000 y( conc ) = (1) 1 + e -b(lnconc-a) rpm) and then by vacuum for 8 min. For GC-MS analyses, the Where 'y' is the fraction of dead animals (dimensionless), (conc) pesticides were eluted from the cartridges with two times 6 ml is the applied dose in µg l-1 on basis of the nominal concentrations ethyl acetate for 6 ml cartridges and 3 ml ethyl acetate for 3 ml at t = 0 and the parameters are: a (ln LC ), b (slope in l µg-1 ) and c cartridges. These samples were either diluted with acetone : 50 (fraction of background effect or mortality in the control). The cyclohexane (1:9) or concentrated to approximately 0.1 ml with parameters were provided by the Genstat program. nitrogen gas and filled up with acetone : cyclohexane (1:9) to a final volume of 1 ml. In a parallel extraction step for LC-PDA From the bioassay results, box plots were created using analysis with the SPE cartridges (200 mg ENV+/6 ml Isolute), statistical program PASW (SPSS version 18). To detect spatial elution was performed with 2 x 5 ml of methanol. This extracts variation in effects among the different sampling sites, a one-way were evaporated at 30-350ºC under nitrogen current to 0.05 ml, ANOVAwith a LSD post-hoc test was conducted in PASW. and reconstituted in 1 ml methanol / water (40:60) mixture. Comparison of temperate versus tropical species The GC-MS extracts were analyzed by gas sensitivities : For comparison of sensitivity of temperate versus chromatography with mass spectrometer with an Agilent 7890A tropical species to ethoprophos and chlorpyrifos, species GC and 5975C MS (Agilent Technologies, Palo Alto, USA), in sensitivity distribution curves (SSDs) were plotted using synchronous SIM Scan mode, a CTC Combipal autosampler previouslyreportedgeometricmeanLC50 valuesofthetemperate (CTC Analytics AG, Switzerland), and a capillary column BP35 speciesrainbowtroutOncorhynchus mykiss and D. magna. A (Agilent Technologies, Palo Alto, USA) (25 m x 0.25 mm x 0.25 temperate and tropical SSD for fish for chlorpyrifos was plotted µm). The data acquisition was carried out using Chemstation usingtheETX2.0software(VanVlaardingen2004)thatusesthe software and NIST05 Mass Spectral Database. The temperature method ofAldenberg and Jawarska (2000) to generate the SSD -1 program was 90ºC for (1 min) to 210ºC with 20ºC min , and then distribution, the derived hazard concentrations HC5 (hazard -1 up to 300ºC with 4ºC min , the interface was held on 280ºC, the concentrationthataffects5%ofthespecies),HC50 andtheir95% ion source on 230ºC, the injector on 230ºC.The sample (1 µl) was lower and upper confident limits. The SSD is described by a injected in splitless mode. Pesticides concentrations were statistical cumulative frequency distribution of either LC50 or quantified by external calibration. NOEC values, based on a log normal distribution.Acute toxicity

data (LC50 , 96 hr) from the EPAdatabase (www.epa.gov/ecotox)

LC-PDA analyses were performed using a Shimadzu wereusedheretoconstructtheSSDs.IfmultipleLC50 datawere HPLC LC-10AD with an SPD-M10A diode array detector available for one species, the geometric mean was taken. Since (Shimadzu, Kyoto, Japan). The chromatographic column was a only two species were available between 23.5N-23.5S LiChroCART HPLC RP-18e column (125 mm x 3 mm x 5 µm (classification for tropical species by (Kwoket al., 2007)), all particle size, Merck, Germany). 50 µl of extracts were analyzed. species between 35N and 35S, with at least one distribution limit The mobile phase consisted of 20 mM sodium acetate in ultra within the tropical region, were used. Other species outside this pure water/methanol 56:44 (solvent A) and methanol (solvent B). range were categorised as temperate. Overlap was allowed Mobile phase was delivered at a flow rate of 0.5 ml min-1 . Gradient outside the range of 23.5N-23.5S. Using these criteria, 12 elution program started from 100% of solvent A, decreased to tropical and 13 temperate fish species were available to 50% A in 15 min and held for 5 min; decreased to 20% A in 5 min construct SSDs for chlorpyrifos. For ethoprophos, only seven and held for 5 min and finally increased to 100% A in 5 min and differentspecieswereavailableofwhichnonecouldbeclassified held 5 min. The total run time was 45 min. Identification was as tropical and only three as temperate according the criteria performed using retention time and the UV-spectra of the above, therefore no SSD was made. The temperate and tropical pesticides included in the analysis. Quantification was performed distributions were compared on significant slope differences by the external standard method. Good recovery yields were using a Kolmogorov-Smirnov test (sign. level 0.05) (PASW obtained for all pesticides between 77Online and 105%. Copy statistics18).

Data evaluation : Tocalculate LC50 and LC 10 values and their 95% Results and Discussion confidence limits, log-logistic analyses were performed using the statistical software Genstat, 12th edition (LawesAgricultural Trust, The analytical recovery for chlorpyrifos was 99.8% and

VSN International Ltd.). The LC50 concentrations were converted for ethoprophos 97.9%. In the tests performed with D. ambigua, from µg l-1 into molarity (mol l -1 ) by using molecular weight hardly any dissipation of the pesticides was observed. This was in 78 N.J. Diepens et al. contrast to the tests performed withP. dovii , where dissipation (Persooneet al., 1985, cited in (Varó et al., 2000)). The difference was up to 99.3%.The fish systems were aerated for 96 hr, which in sensitivity may be the result of differences in enzymatic could have led to high volatilization of both chemicals during the expression or metabolism (Boone and Chambers, 1997). test since both chemicals have a relatively high Henry's law Furthermore, arthropods crustaceans are usually the most constant (0.478 Pa m3 mol -1 at 25ºC for chlorpyrifos and 0.0135 Pa sensitive taxonomic group to insecticides, i.e. compared to m3 mol -1 at 25ºC for ethoprophos (Tomlin,2003). For both species, vertebrates and other aquatic groups (Maltbyet al., 2005; Daam nominal values were used for further calculations. and Van den Brink 2010). The variation in toxicity between two compounds could also be explained by difference in chemical All mortality levels in the controls were below 20% structure leading to different physico-chemical properties such as (maximum acceptable level set by the (OECD, 1992) forDaphnia ) solubility and hydrophobicity and thus to difference in uptake and except for the test withD. ambigua and ethoprophos,in which elimination of the compound by the individuals. 33% mortality was observed in the control. These test results are used for further calculations because they still yielded a clear ForP.dovii no toxicity data have previously been reported dose response relationship but the results should be interpreted for both OP compounds. The LC50 value ofP.dovii reported in this with caution. paper for ethoprophos is similar to the earlier reported LC50 values -1 forsamespeciesandcompound(96hrLC50 :54,158and220µgl ). Results of acute toxicity tests showed that D. ambigua The fish used in these experiments were older and therefore was more sensitive to both organophosphates (OPs) compared possibly less sensitive, which could explain the slightly higher toP. dovii and that chlorpyrifos was more toxic to these two LC50 values. Compared to the standard temperate test species on -1 species than ethoprophos, although they have a similar mode of O. mykiss (geometric mean 96 hr LC50 was 19 µg l for action (i.e. both are neurotoxins that cause inhibition of AChE) chlorpyrifos (n=13) and 3153 µg l-1 for ethoprophos (n=6)), P.dovii (Table 1 and Fig. 2). The difference in toxicity of two pesticides was less sensitive for chlorpyrifos and more sensitive to and in species sensitivity have been reported earlier for other OPs ethoprophos, both within an order of magnitude.

Dose response curveParachromis dovii chlorpyrifos Dose response curveParachromis dovii ethoprophos 100 100 (A) Measured mortality 80 80 Measured mortality (B) 60 60

40 40

Mortality (%) Mortality (%) 20 20

0 0 1 10 100 1000 1 10 100 1000 Log concentration Log concentration

Dose response curveDaphnia ambigua chlorpyrifos Dose response curve Daphnia ambigua ethoprophos 100 100 Measured mortality Measured mortality 80 80 (C) (D) 60 60

40 40

Mortality (%)

Mortality (%) 20 Online20 Copy 0 0 1 10 100 1000 1 10 100 1000 Log concentration (m g l-1 ) Log concentration (m g l-1 ) Fig. 2 : Dose response curves forParachromis dovii (A and B) and Daphnia ambigua (C and D) and for chlorpyrifos (A and C) and ethoprophos (B and D) (n=3 with 10 individuals per replicate) Effect of pesticides on aquatic ecosystems in Costa Rica 79

Table 1 : Results of acute toxicity tests for chlorpyrifos and ethoprophos withParachromis dovii and Daphnia ambigua

-1 -1 -1 LC10 (µg l ) (95% CI) LC50 (µg l ) (95% CI) LC50 (mol l ) Parachromis dovii Chlorpyrifos 53 (29, 98) 117 (87, 158) 3.3*10-7 Ethoprophos 73 (43, 125) 242 (179, 328) 1.0*10-6 Daphnia ambigua Chlorpyrifos 0.021(0.013, 0.034) 0.039 (0.029, 0.053) 1.1*10-10 Ethoprophos 7.01 (6.66, 7.37) 7.63 (7.30, 7.97) 3.1.10-8

The SSD (Fig. 3) shows that the LC50 ofP.dovii falls within The LC50 ofD. ambigua for chlorpyrifos was higher than -1 the confidence limits (95%) of tropical HC50 but outside the range the earlier published LC50 value of 0.035 µg l (95% CI 0.032- of the temperate. Based on this SSD, tropical fish species were 0.037 µg l-1 ) (Harmonet al., 2003). Ethoprophos showed a steep indicated as less sensitive to chlorpyrifos, yet no statistical dose response curve, which indicates that a small change in difference was detected between the temperate and tropical concentration can result in a large change in the mortality distributions (p=0.975). This is consistent with a previous response (Fig. 2D). This was illustrated by LC10 value that was comparison study for fish performed by Dyeret al. (1997) who very close to the LC50 value (Table 1).D. ambigua was 10 to 100 found no difference between tropical and temperate sensitivity for times more sensitive than the standard temperate test species, D. -1 carbaryl, lindane, malathion, pentachlorophenol and phenol. magna (geometric mean of LC50 : 0.8 µg l for chlorpyrifos (n=2) -1 Moreover, Ricoet al. (2010) (methylparathion) and Rico et al. and 63.9 µg l for ethoprophos (n=2)).D. ambigua LC50 values for (2011) (malathion and carbendazim) showed that three chlorpyrifos were found lowest amongall the arthropods and fish indigenous tropical fishes and freshwater invertebrates species species in de EPAdatabase. Tropical HC5 for chlorpyrifos derived of the Amazon were not more sensitive than their temperate by Maltbyet al. (2005) from an SSD with acute toxicity data for counterparts. arthropods had a value of 0.06 µg l-1 (95% CI 0.002-0.16) and was not statically different from the temperate value (0.13 µg l-1 ). The Castilloet al. (1997) recommended the guapote fish as a lower confidence level of HC5 derived from an SSD gives a tropical freshwater standard test species because of its economic conservative estimate of the ecosystem threshold concentrations value and/or sensitivity. Its economic value is considerable since (Maltbyet al., 2005). With a LC50 value below the tropical HC5 , D. it is popular in local fisheries and sport fishing (tourism) and is ambigua is a sensitive species, therefore they could be used as a used for human consumption. In addition, this fish species is easy tropical freshwater indicator species instead ofD. magna in to cultivated, lay a substantial number of eggs (800-1000) and is tropical risk assessments. Furthermore, this species is easy to therefore suitable for toxicity testing. However, the results show culture and is suitable for routine toxicity tests (Harmon et al., that the guapote fish is not particularly sensitive for the tested 2003). Moreover,Daphnia sp. are taxonomic stable, sensitive to compounds. Therefore, in order to protect 95% of tropical fish wide range of chemicals and widely distributed (Sarma and species from chlorpyrifos, an uncertainty factor of 100 should be Nandini, 2006). applied to its LC50 to obtain a threshold value that lies within the confident limits of tropical HC5 (Fig. 3). More toxicity tests need to The physico-chemical parameters of the surface water be conducted to reveal its sensitivity to other toxicants. measured at the field sites showed natural fluctuations in pH and conductivity; i.e. higher pH and conductivity levels towards the sea. In the tropics, it should be taken into account that warmer water can contain less oxygen. In Canal Pama, the lowest oxygen -1 1.0 SSD fish chlorpyrifos concentration (1.84 mg l ) was measured, dead fish were observed and a change in water colour (from brown to green) was 0.8 observed within few hours, indicating a rapid algal bloom. With five different pesticides (difenoconazole, epoxiconazole, ethoprophos, 0.6 fenamiphos, triadimenol) detected, second highest number of Temperate: HC =0.66 5 pesticides were detected in Canal Pama (Table 2). A few days 0.4 (0.10-2.09) HC50 = 16.29 (6.37-41,68) before the analysis it had rained heavily (local meteorological Tropical: HC = 1.36 data), probably causing runoff of pesticides and nutrients from 0.2 Online5 Copy (0.11- 6.05) HC50 = many banana plantations that drained into Canal Pama and at this

Potentially affected fraction Potentially affected 75.30 (22.00-257.77) time fish mortality was observed. The highest concentration was 0.0 -1 found for the nematicide fenamiphos (2 µg l )(Table 2), which was 0.01 0.1 1 10 100 1000 10000 100000 -1 lower than the acute 96 hr LC50 for fish, crustacean and algae but LC50 (m g l ) slightly higher than 48 hr LC50 for aquatic invertebrates (EU Fig. 3 : Temperate and tropical species sensitivity distribution (SSD) for FOOTPRINT).The death of fish at the time of sampling (a few days chlorpyrifos and fish after the first fish mortality) was probably caused due to 80 N.J. Diepens et al.

Table 2 : Concentrations (µg l-1 ) of different pesticides per sample site found during one sampling time

Pesticides RMD control Bridge RMD CañoAzul Canal Pama RMD X Freeman LMD X RMD LMD mid LMD X Pacuare Ametryna 0.5 traces 0.06 0.06 Carbaryl 0.4 Carbofuran 5.0 Chlorpyrifos traces Chlorothalonil 0.05 traces 0.07 0.05 Diazinon 0.05 Difenoconazol 0.8 0.3 Diuron 2.0 0.1 0.1 0.2 Epoxiconazol 0.6 0.6 0.1 0.2 0.3 0.2 Ethoprofos 0.1 traces 0.9 0.9 1.0 Fenamiphos 2.0 Metalaxyl traces Triadimenol 0.07 combination of multiple stressors i.e.; low oxygen concentration Castilloet al. (1997). Earlier sampling in the catchment showed combined with pesticide exposure (Fig. 4). that 67% of the samples taken from RMD, the following seven pesticides were frequently found in water and/or sediment: Most pesticides (ametryn, carbofuran, carbaryl, chlorothalonil, bromacil, difenoconazole, diuron, ethoprophos, chlorpyrifos, diuron, difenoconazole, epoxiconazole, esfenvalerate and cypermethrin. In this study, bromacil, ethoprophos, fenamiphos) were found in Caño Azul (Table 2) as esfenvalerate and cypermethrin were not found in any of the was hypothesized due to the extensive drainage of banana and samples. In general, epoxiconazole and ethoprophos were most pineapple plantations but no dead fish or algae blooms were -1 -1 frequently detected at respectively six and five of the eight observed here. Carbofuran (5.0 µg l ) and diuron (2.0 µg l ) were samples, while only traces of chlorpyrifos were detected at Caño found at high concentrations. At the control site of Rio Madre de Azul. This reflects the wide use of ethoprophos and its mobility Dios (RDM control) no pesticides were found confirming its through the watershed, possible due to its high solubility (700 mg suitability as use as a control (Table 2). From the CañoAzul to the l-1 at 20ºC (Tomlin, 2003)). Robinsonet al. (1999) found that 30% crossing of RMD with Freeman (RMD X Freeman), the of the applied ethoprophos runs off into rivers while the runoff concentration and the number decrease from eight pesticides to potential of chlorpyrifos is low (Ware and Racke, 1993). Although one. The decrease of pesticides can be explained by biological chlorpyrifos is widely used in the fruit protection bags, its leaching and chemical degradation, such as photolytic degradation, potential due to rain events could be minimal because of its low vaporization and sorption to organic matter. Compared to solubility (1.4 mg l-1 at 25ºC (Tomlin, 2003)) and because the temperate systems, these processes can be faster due to the product is the applied to the inside of the bag, thus with limited tropical conditions (see (Daam and Van den Brink 2010) for a exposure to rain. When reaching the water column, the review), though many of the detected pesticides are known to be compound likely to bind to organic material in the sediment or persistent in water. In lagoon, the amount and concentration of living animals and may bioaccumulate in the biota (Varóet al. , the pesticides increased probably due to entry of Canal Pama and 2000; Varóet al. , 2002). On the other hand, it is rapidly other side streams. Especially, ethoprophos (0.9 - 1.0 µg l-1 ) and metabolized into less toxic molecules and eliminated from the epoxiconazole (0.2 - 0.3 µg l-1 ) has been found consistently body (Ware and Racke, 1993; Barron and Woodburn 1995). More throughout the whole lagoon.It should be noted that one single research is required to determine the fate of chlorpyrifos in the pesticide and physico-chemical parameter analysis is not plastic banana protection bags. sufficient to obtain a full overview of pesticide concentration dynamics due to the spatial and temporal variation in applications In Central America, maximum environmental and rain events. Currently, a long-term monitoring program (by concentrations found in water, sediment and biota were higher for the TROPICA project and IRET) is running, which aims at the chlorpyrifos (water 0.18 µg l-1 ; sediment 34.2 µg kg -1 ; biota 8 µg determination of pesticide concentrations in water and sediment kg-1 (Castilloet al ., 1997)) than for ethoprophos (water 0.10 µg l-1 in this watershed and evaluates the overallOnline risks of pesticides to (Unpublished Copy data); sediment 1.98 µg kg-1 (Robinsonet al ., native species in this ecosystem.This will give important insight to 1999). Ethoprophos concentrations measured in this study (max the dynamics in such systems. concentration in water 1µg l-1 ) were 10 fold higher than previous reported concentration while chlorpyrifos concentrations were In general, the most frequently found pesticide groups in under the detection limit in this study. Calculating the risk quotient this study were fungicides followed by herbicides and insecticides (maximal environmental concentration/predicted no effect

/ nematicides (Table2), which is in accordance with the findings of concentration for most sensitive species in this study (LC50 Effect of pesticides on aquatic ecosystems in Costa Rica 81

Pesticide Pesticides Death fish application Low dissolved oxygen Death algea Decaying organic concentrations Rain event Runoff grazers material

Nutrients Algae bloom Higher water Sun temperature

Fig. 4 : Conceptual model of environmental processes possibly leading to fish kills after a rain event

BioassayParachromis dovii 72 hr BioassayDaphnia magna 48 hr a a c c 40 40 (A) (B)

a 30 30

b ba 20 20

Mortality (%) Mortality (%)

a a 10 10 a

0 0 Negative Control Bridge Canal Lagoon Negative Control Bridge Canal Lagoon control RMD RMD Pama RMD control RMD RMD Pama RMD Fig. 5 : Bioassays for (A)P.dovii (72 hr) and (B) D. magna (48 hr) showing mortality per sample site in the watershed Rio Madre de Dios. Values are mean of three replicates+ SD. Differences in letters indicate sites with significant difference (a =0.05) divided by a safety factor 10) gives a rough estimate that there is a Therefore, it proved to be a good indicator of pollution and is substantial risk in tropical ecosystems for both compounds with useful for rapid detection of environmental contamination. The risk quotients of 16 and 46 for ethoprophos and chlorpyrifos toxicity test, showed thatD. ambigua was more sensitive than D. respectively. Besides acute effects, chronic effects may occur at magna for the evaluated compounds. Since it is an indigenous lower concentrations. tropical species, it might be even more suitable as an indicator species in the tropics and replaceD. magna in a tropical test Fig. 5 shows the difference in response of the bioassays battery. performed withP. dovii and D. magna among the sampled sites. Mortality of both species was below 40% at all sampling sites. For In the laboratory toxicity tests with different concentrations fish, no significant difference (p=0.482) in mortality between sites of ethoprophos and chlorpyrifos, there was a significant treatment could be detected. Significant differences in mortality among sites effect on ChE activity analysed in whole fish tissue (Fig. 6Aand 6B) (p=0.001) were found forD. magna . The LSD post-hoc test (one-way ANOVA). Post-hoc comparisons showed that for showed that the observed mortality at Bridge RMD was ethoprophos the three highest concentrations were not significantly different from Canal Pama (p=0.001) and lagoon Onlinesignificantly Copy different, and for chlorpyrifos 10 and 30 µg l-1, and the (p=0.007) but similar to both controls. RMD and Lagoon RMD highest two concentrations were not significantly different (Fig. 6). were not significantly different (p=0.290) but differed in observed -1 mortality with all other sites. The unexpected high fish mortality rate (33%) at 3 µg l chlorpyrifos concentration (Fig. 2A) is probably due to other factors, such as ForD. magna a higher mortality in the most polluted sites biological effects or a fungal infection, than the pesticides toxicity and difference among sample sites could be detected (Fig. 5). since high ChE activity does not reveal any inhibition. Fig. 6 shows 82 N.J. Diepens et al.

(A) (B) 150.00 a 250.00

a -1 -1 b -1200.00 -1

100.00 150.00

c 100.00 50.00 c b d 50.00 c

ChE activity ( n mol mg protein min ) d ChE activity ( n mol mg protein min ) d d d 0.00 0.00 0 3 10 30 100 300 0 20 60 180 540 1620 Concentration (m g l-1 ) Concentration (m g l-1 ) Fig. 6 : ChE inhibition in P.dovii after 96 hr exposure to (A) chlorpyrifos and (B) ethoprophos. Values are mean of three replicates+ SD. Difference in letters indicate that the treatments were significantly different from each other (alpha =0.05)

effects at individual level, such as changes in behaviour, ChE activity in bioassays 80.00 reproduction and feeding success, which could affect populations ba in long term.

-1 -1 ChE activities measured in the bioassay (Fig. 7) shows 60.00 a cd that ChE activity measured in the negative control of the toxicity tests was significantly similar to ChE activity levels measured at 40.00 the field control site (RMD control). ChE activity in fish exposed to dc water from Bridge RMD and Canal Pama did not differ from each d other and were both significantly lower than control. However, 20.00 Canal Pama was similar to Lagoon RMD. The strongest ChE

Activity ( n mol mg protein min ) inhibition was found in the Lagoon RMD, where only one OPcould be detected. This was surprising since there was also a higher 0.00 dilution in the lagoon with incoming seawater from the Carribean Negative RMD Bridge Canal Lagoon Sea and from other adjacent streams. These other streams, control Control RMD Pama RMD however, could bring other OPs and carbamates that we did not Fig. 7 : ChE inhibition inP. dovii after 72 hr exposure to surface water look for in our chemical analyses. During the bioassays, no fish taken from watershed Rio Madre de Dios. Values are mean of three mortality was observed, the ChE inhibition however, showed a replicates+ SD. Differences in letters indicate that the treatments were clear response i.e. 37% inhibition in Canal Pama and 62% significantly different from each other (alpha =0.05). inhibition in the Lagoon RMD. This again indicates the usability of this biomarker as an early warning indicator of exposure to OPs that the amount of ChE decreased with increasing concentrations and carbamates. Yet, the majority of the pesticides found with significant difference for ethoprophos between the control and consisted of fungicides and herbicides, which have a different the first concentration and for chlorpyrifos between the first and mode of action than OPs and carbamates. Therefore, a combination with other biomarkers and/or toxicity tests is needed second concentration. This and the threshold value of 20% in order to detect pesticide contamination and asses its inhibition (Varóet al., 2007) indicates that inhibition begins for environmental risks in this watershed. ethoprophos at 20 µg l-1 (32%) and for chlorpyrifos at 10 µg l-1 (38%). For both pesticides no immobileOnline or dead fish were observed InCopy conclusion, the proposed tropical species were similar at these concentrations while ChE inhibition showed already sub or more sensitive than their temperate counterparts. More toxicity lethal effects. Therefore, this biomarker could be used as an early data are needed about to evaluate possible similarities or warning indicator. Since ChE recovery was low or even assumed respectively differences between temperate and tropical species to be irreversible it can provide long-term exposure information to chemical contamination. Further studies are also needed to instead of a snapshot by chemical analysis of water samples. investigate how protective safe environmental concentrations,

Moreover, inhibition of ChE might be linked to different indirect derived with LC50 s of temperate species, are for protecting tropical Effect of pesticides on aquatic ecosystems in Costa Rica 83 aquatic natural communities. Both recommended tropical and Sediments. Burlington, Ontario, National Water Research species were found suitable for toxicity testing, and might be used Institute (NWRI), Environmental Canada. (1989). as a part of a tropical test battery and environmental risk Dyer, S.D., S.E. Belanger and G.J. Carr:An initial evaluation of the use of assessment in tropical regions. Euro/NorthAmerican fish species for tropical effects assessments. Chemosphere,35 , 2767-2781 (1997). Acknowledgment Ellman, G.L., K. D. Courtney, V. Andres jr and R.M. Featherstone: A new and rapid colorimetric determination of acetylcholinesterase This study was funded by the TROPICA project, Swedish activity. Biochemical Pharmacology,7 , 88-95 (1961). Research Council FORMAS, grant nr 2005-473-3035-21 and by EU FOOTPRINT: database www.eu-footprint.org. Retrieved 24-02-2010 (2010). complementary funding from IRET. We are grateful for the Harmon, S.M., W.L. Specht and G.T. Chandler: A comparison of the valuable help from all people from both the ECOTOX and LAREP daphnidsCeriodaphnia dubia and Daphnia ambigua for their laboratories of IRET, Universidad Nacional, Heredia, Costa Rica. utilization in routine toxicity testing in the Southeastern United We also want to thank Julio Knight and his family for assistance States.Archi. Environ. Contam.Toxicol., 45 , 79-85 (2003). during the fieldwork and important knowledge of the study area. Kwok, K.W.H., K.M.Y. Leung, G.S.G. Lui, V.K.H. Chu, P.K.S. Lam, D. Morritt, L. Maltby, T.C.M. Brock, P.J. Van den Brink, M.S.J. Warne References and M. Crane: Comparison of tropical and temperate freshwater animal species' acute sensitivities to chemicals: Implications for Aldenberg, T. and J.S. Jaworska: Uncertainty of the hazardous deriving safe extrapolation factors. Integra. Environ. Asse. concentration and fraction affected for normal species sensitivity Manage., 3, 49-67 (2007). sistributions.Ecotoxicol. Environ. Saf., 46 , 1-18 (2000). Lacher, T.E. and M.I. Goldstein: Tropical ecotoxicology: Status and Barraza, D., K. Jansen, B.V. de Joode and C. Wesseling: Pesticide use needs.Environ.Toxicol.Chem., 16 , 100-111 (1997). in banana and plantain production and risk perception among local Lopes, M.C., A.P. Carvalho, L. Guilhermino and A.M.V.M. Soares: actors in Talamanca, Costa Rica.Environ. Res., 111 , 708-717. Inhibition of acetylcholinesterase activity as effect criterion in acute (2011). tests with juvenileDaphnia magna . Chemosphere ,32 , 727-738. Barron, M.G. and K.B. Woodburn: Ecotoxicology of chlorpyrifos. Revi. (1996). Environ. Contami.Toxicol.,144, 1-93 (1995). Maltby, L., N. Blake, T.C.M. Brock and P.J. Van den Brink: Insecticide Boone, J.S. and J.E. Chambers: Biochemical factors contributing to species sensitivity distributions: Importance of test species toxicity differences among chlorpyrifos, parathion, and methyl selection and relevance to aquatic ecosystems. Environ. Toxicol. parathion in mosquito fish (Gambusia affinis ). Aqu. Toxicol., 39 , Chem., 24, 379-388 (2005). 333-343 (1997). Matlock Jr, R.B. and R. de la Cruz:An inventory of parasitic Hymenoptera Bradford, M.M.: A rapid and sensitive method for the quantitation of in banana plantations under two pesticide regimes. Agricu. microgram quantities of protein utilizing the principle of protein-dye Ecosys. Environ.,93 , 147-164 (2002). binding.Analytical Biochem., 72 , 248-254 (1976). Mortensen, S.R., K.A. Johnson, C.P. Weisskopf, M.J. Hooper, T.E. Bravo, D.V., E. de la Cruz, M. Herrera Ledezma and F. Ramírez: Lacher and R.J. Kendall: Avian exposure to pesticides in Costa Agriculture pesticides use as a tool for monitoring health hazards Rican banana plantations,60 , 562-568 (1998). (In Spanish). Uniciencia,27 , 351-376 (2013). OECD: Guidance document for aquatic effects assessment. Paris, Castillo, L.E., E. De laCruz and C. Ruepert: Ecotoxicology and pesticides France.(1992). in tropical aquatic ecosystems of Central America. Environ. Pais, E.: http://wvw.nacion.com/L-1n_ee/2009/febrero/25/pais 1886110. Toxicol.Chem., 16, 41-51 (1997). html. Retrieved 27-02-2010. (2009). Castillo, L.E., E. Martinez, C. Ruepert, C. Savage, M. Gilek, M. Pinnock Pais, E.: http://www.elpais.cr/articulos.php?id=19296. Retrieved 27-02- and E. Solis: Water quality and macroinvertebrate community 2010 (2010). response following pesticide applications in a banana plantation, Pringle, C.M. and A. Ramírez: Use of both benthic and drift sampling Limon, Costa Rica.Sci. TotalEnviron., 367 , 418-432 (2006). techniques to assess tropical stream invertebrate communities Castillo, L.E., C. Ruepert, F. Ramírez, B. Van Wendel de Joode, V. Bravo along an altitudinal gradient, Costa Rica.Freshwater Biol., 39 , and E. De la Cruz: Plaguicidas y otros contaminantes. Décimo 359-373 (1998). Octavo Informe Estado de La Nación en Desarrollo Humano Ramírez, F.: Importación de plaguicidas en Costa Rica: Periodo 2007- Sostenible. San José, Costa Rica, 32 (2012). 2009 Heredia, Costa Rica, Informe elaborado para el Proyecto Castillo, L.E., C. Ruepert and E. Solis: Pesticide residues in the aquatic Reduciendo el Escurrimiento de Plaguicidas al Mar Caribe environment of banana plantation areas in the north Atlantic zone (REPCar), Universidad Nacional (2010). of Costa Rica.Environ.Toxicol.Chem., 19 , 1942-1950 (2000). Ramírez, F., F. Chaverri, E. de la Cruz, C. Wesseling, L. Castillo and V. Daam, M.A., S.J.H. Crum, P.J. Van den Brink and A.J. A. Nogueira: Fate Bravo: Importación de plaguicidas en Costa Rica, Periodo 1977- and effects of the insecticide chlorpyrifosOnline in outdoor plankton- 2006. Copy Serie Informes Técnicos IRET No.6. Heredia, Costa Rica, dominated microcosms in Thailand.Environ. Toxicol. Chem .,27 , Instituto Regional de Estudios en Sustancias Tóxicas, Universidad 2530-2538 (2008). Nacional,58 (2009). Daam, M. A. and P.J. Van den Brink: Implications of differences between REPCar: Experiencias exitosas para educer el impacto de la agricultura temperate and tropical freshwater ecosystems for the ecological sobre los ecosistemas costeros. Resumen de resultados y logros., risk assessment of pesticides.Ecotoxicol., 19 , 24-37 (2010). ReduciendoElEscurrimientoDePlaguicidasAlMarCaribe.(2011). Dutka, B.J.: Methods for Toxicological Analysis of Waters, Wastewaters Rico, A., R. Geber-Corrêa, P. S. Campos, M. V. Garcia, A. V. Waichman 84 N.J. Diepens et al.

and P. J. Van den Brink: Effect of parathion-methyl on Amazonian distribution based hazardous concentration and fraction affected. fish and freshwater invertebrates: A comparison of sensitivity with RIVM. Bilthoven,The Netherlands (2004). temperate data.58 , 765-771 (2010). Varó, I., J.C. Navarro, B. Nunes and L. Guilhermino: Effects of dichlorvos Rico,A.,A.V.Waichman, R. Geber-Corrêa and P.J.van den Brink: Effects aquaculture treatments on selected biomarkers of gilthead sea bream of malathion and carbendazim on Amazonian freshwater (Sparus aurata L.) fingerlings. Aquaculture, 266 , 87-96 (2007). organisms: comparison of tropical and temperate species Varó, I., R. Serrano, E. Pitarch, F.Amat, F.J. Lopez and J.C. Navarro: Bio sensitivity distributions,Ecotoxicology ,20 , 625-634 (2011). accumulation of chlorpyrifos through an experimental food chain: Robinson, D.E., A. Mansingh and T.P. Das Gupta: Fate and transport of Study of protein HSP70 as bio marker of sublethal stress in fish. ethoprophos in the Jamaican environment. Sci. Total Environ., Arch. Environ. Contami.Toxicol., 42, 229-235 (2002). 237–238, 373-378 (1999). Varó, I., R. Serrano, E. Pitarch, F. Amat, F.J. López and J.C. Navarro: Sarma, S.S.S. and S. Nandini: Review of recent ecotoxicological studies Toxicityand bio concentration of chlorpyrifos in aquatic organisms: on cladocerans.J. Environ. Sci. Hlth., 41 , 1417-1430. (2006). Artemia parthenogenetica(Crustacea), Gambusia affinis and Schroer, A.F.W., J.D.M. Belgers, C.M. Brock, A.M. Matser, S.J. Maund and P.J. Brink: Comparison of laboratory single species and field Aphanius iberus(Pisces), Bull. Environ. Contami. Toxicol., 65 , population-level effects of the pyrethroid insecticide λ-cyhalothrin 623-630 (2000). on freshwater invertebrates.46 , 324-335. (2004). Ware, G. and K. Racke: Environmental fate of chlorpyrifos. Revi. Thompson, H.: Esterases as markers of exposure to organophosphates Environ. Contami. Toxicol., Springer, New York.131 , 1-150 and carbamates.8 , 369-384 (1999). (1993). Tomlin, C.D.S.: The pesticide manual a world compendium. UK, BCPC. Waylen, P.R., M.E. Quesada and C.N. Caviedes: Temporal and spatial (2003). variability of annual precipitation in Costa Rica and the Southern Van Vlaardingen, P., T.P. Traas and T. Aldenberg: ETX2.0 Normal Oscillation.Inter. J. Climatol., 16 , 173-193. (1996).

Online Copy